{
  "id": "image-mathematics",
  "title": "Image (mathematics)",
  "lang": "en",
  "source_lang": "en",
  "summary": "The image of a function is the set of outputs it attains, either on its entire domain or on a specified subset.",
  "category": "mathematics",
  "subcategory": "algebra",
  "type": "concept",
  "aliases": [
    "Direct image",
    "Forward image"
  ],
  "generated_at": "2026-11-05",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/image-mathematics/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/image-mathematics/",
      "json": "https://aiwiki.page/zh/image-mathematics.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "In mathematics, the image of a function is the set of values it actually takes. More generally, the image of a subset of the function’s domain consists of the outputs obtained from inputs in that subset. For a function \\(f:X\\to Y\\) and a subset \\(A\\subseteq X\\), the image is\n\n\\[f[A]=\\{f(x):x\\in A\\}.\\]\n\nThe image of the entire domain, \\(f[X]\\), is also written \\(\\operatorname{im}f\\). It is a subset of the codomain \\(Y\\), but need not equal it. Images are used throughout mathematics to describe attainable values, transformed sets, and the outputs of structure-preserving maps."
    },
    {
      "heading": "Definition and terminology",
      "text": "Let \\(f:X\\to Y\\) be a function with domain \\(X\\) and codomain \\(Y\\). For every subset \\(A\\subseteq X\\),\n\n\\[f[A]=\\{y\\in Y:\\text{there exists }x\\in A\\text{ such that }f(x)=y\\}.\\]\n\nSeveral related expressions must be distinguished:\n\n- Image of an element: \\(f(x)\\), an element of \\(Y\\).\n- Image of a subset: \\(f[A]\\), a subset of \\(Y\\).\n- Image of the function: \\(f[X]\\), the set of all outputs attained by \\(f\\).\n\nThe notation \\(f(A)\\) is also commonly used for the image of a set. Square brackets help distinguish applying a function to an element from taking the image of a collection of elements.\n\nThe word range often means the image of the whole domain, but some authors use it for the codomain. Explicitly distinguishing image from codomain avoids this ambiguity. A function is surjective precisely when\n\n\\[f[X]=Y.\\]\n\nAn image records which values occur, not how many times they occur. Distinct inputs producing the same output contribute only one element to the image."
    },
    {
      "heading": "Examples",
      "text": "A real-valued square function\n\nFor the function\n\n\\[f:\\mathbb R\\to\\mathbb R,\\qquad f(x)=x^2,\\]\n\nwhere \\(\\mathbb R\\) is the set of real numbers,\n\n\\[\\operatorname{im}f=[0,\\infty).\\]\n\nIts codomain contains negative numbers, but its image does not.\n\nThe image depends on which inputs are allowed:\n\n\\[f-2,1=[0,4], \\qquad f[(-2,1)]=[0,4).\\]\n\nThe second image includes \\(0\\), because \\(0\\) belongs to the input interval, but excludes \\(4\\), because neither \\(2\\) nor \\(-2\\) belongs to that interval.\n\nA parametrized curve\n\nConsider\n\n\\[g:\\mathbb R\\to\\mathbb R^2,\\qquad g(t)=(\\cos t,\\sin t).\\]\n\nIts image is the unit circle:\n\n\\[g[\\mathbb R]=\\{(u,v)\\in\\mathbb R^2:u^2+v^2=1\\}.\\]\n\nThe image is a geometric set; it does not retain information about the direction, speed, or repeated traversal specified by the parametrization.\n\nImage versus graph\n\nThe graph of a function \\(f:X\\to Y\\) is\n\n\\[\\{(x,f(x)):x\\in X\\}\\subseteq X\\times Y.\\]\n\nIt records input–output pairs. By contrast, the image is a subset of \\(Y\\) containing only outputs. Thus the graph of \\(x\\mapsto x^2\\) is a parabola, whereas its image as a real-valued function is \\([0,\\infty)\\)."
    },
    {
      "heading": "Images and inverse images",
      "text": "For a subset \\(B\\subseteq Y\\), the inverse image, or preimage, is\n\n\\[f^{-1}[B]=\\{x\\in X:f(x)\\in B\\}.\\]\n\nAn image moves a set of inputs forward to outputs; an inverse image collects all inputs whose outputs lie in a specified set.\n\nThe notation \\(f^{-1}[B]\\) does not require an inverse function to exist. It is defined for every function, including functions that are neither injective nor surjective.\n\nFor example, if \\(f(x)=x^2\\), then\n\n\\[f^{-1}[\\{4\\}]=\\{-2,2\\}, \\qquad f^{-1}[(-\\infty,0)]=\\varnothing.\\]\n\nThe preimage of a singleton, \\(f^{-1}[\\{y\\}]\\), is called the fiber over \\(y\\). A point \\(y\\) belongs to the image exactly when its fiber is nonempty.\n\nImages and inverse images satisfy\n\n\\[A\\subseteq f^{-1}[f[A]]\\]\n\nand\n\n\\[f[f^{-1}[B]]=B\\cap f[X]\\subseteq B.\\]\n\nThe first inclusion can be strict because an input outside \\(A\\) may share its output with an input inside \\(A\\). It is an equality for every \\(A\\subseteq X\\) exactly when \\(f\\) is injective. The second inclusion is an equality for every \\(B\\subseteq Y\\) exactly when \\(f\\) is surjective.\n\nA useful relationship is\n\n\\[f[A]\\subseteq B \\quad\\Longleftrightarrow\\quad A\\subseteq f^{-1}[B].\\]\n\nIt expresses a basic correspondence between direct and inverse images."
    },
    {
      "heading": "Set-theoretic properties",
      "text": "Taking images defines a function between power sets:\n\n\\[f_*:\\mathcal P(X)\\to\\mathcal P(Y), \\qquad A\\mapsto f[A].\\]\n\nInclusion and unions\n\nImages preserve inclusion:\n\n\\[A\\subseteq C \\quad\\Longrightarrow\\quad f[A]\\subseteq f[C].\\]\n\nThey also preserve arbitrary unions:\n\n\\[f\\left[\\bigcup_{i\\in I}A_i\\right] = \\bigcup_{i\\in I}f[A_i].\\]\n\nIn particular, the image of the empty set is empty.\n\nIntersections\n\nFor any nonempty family of subsets of \\(X\\),\n\n\\[f\\left[\\bigcap_{i\\in I}A_i\\right] \\subseteq \\bigcap_{i\\in I}f[A_i].\\]\n\nEquality need not hold: the same output may be attained in each set by different inputs, without any input belonging to all the sets.\n\nFor example, with \\(f(x)=x^2\\), \\(A=\\{-1\\}\\), and \\(C=\\{1\\}\\),\n\n\\[f[A\\cap C]=\\varnothing, \\qquad f[A]\\cap f[C]=\\{1\\}.\\]\n\nIf \\(f\\) is injective, it preserves all nonempty intersections.\n\nComposition\n\nFor functions \\(f:X\\to Y\\) and \\(g:Y\\to Z\\), composition satisfies\n\n\\[(g\\circ f)[A]=g[f[A]].\\]\n\nConsequently,\n\n\\[\\operatorname{im}(g\\circ f)=g[\\operatorname{im}f].\\]\n\nUnlike direct images, inverse images preserve arbitrary unions, arbitrary intersections, and complements relative to the specified domain and codomain. This distinction is central to their use in topology and measure theory."
    },
    {
      "heading": "Factorization through the image",
      "text": "Every function \\(f:X\\to Y\\) can be factored as\n\n\\[X\\xrightarrow{\\ \\widetilde f\\ }f[X] \\xrightarrow{\\ i\\ }Y,\\]\n\nwhere \\(\\widetilde f(x)=f(x)\\) is surjective and \\(i\\) is the inclusion map, which is injective.\n\nThis factorization separates two features of a function: which inputs have the same output, and where the attained outputs sit inside the codomain.\n\nDefine an equivalence relation on \\(X\\) by\n\n\\[x\\sim x' \\quad\\Longleftrightarrow\\quad f(x)=f(x').\\]\n\nThe map\n\n\\[X/{\\sim}\\ \\longrightarrow f[X], \\qquad [x]\\longmapsto f(x),\\]\n\nis a bijection. Thus the image can be identified with the set obtained by grouping together inputs that have equal outputs. This is a set-theoretic identification; additional algebraic or topological structure requires further conditions."
    },
    {
      "heading": "Images in algebra",
      "text": "For a linear map \\(T:V\\to W\\), its image is a linear subspace of \\(W\\):\n\n\\[\\operatorname{im}T=\\{T(v):v\\in V\\}.\\]\n\nIf \\(T\\) is represented by a matrix, its image is the span of that matrix’s columns. Its dimension is the rank. In finite dimensions, the rank–nullity theorem states\n\n\\[\\dim V=\\dim(\\ker T)+\\dim(\\operatorname{im}T).\\]\n\nFor a group homomorphism \\(h:G\\to H\\), the image is a subgroup of \\(H\\). The first isomorphism theorem identifies it with the quotient group\n\n\\[G/\\ker h\\cong\\operatorname{im}h.\\]\n\nThis gives an algebraic version of grouping inputs according to their common output."
    },
    {
      "heading": "Images in topology and measure theory",
      "text": "For a continuous function, images preserve some important properties of sets:\n\n- The image of a compact space is compact.\n- The image of a connected space is connected.\n\nContinuity alone does not imply that images of open sets are open or that images of closed sets are closed. For example, \\(x\\mapsto x^2\\) maps the open interval \\((-1,1)\\) to \\([0,1)\\), which is not open in \\(\\mathbb R\\).\n\nA function is called an open map if it sends open sets to open sets, and a closed map if it sends closed sets to closed sets. These are properties of direct images, distinct from continuity, which is defined through inverse images of open sets.\n\nA similar distinction occurs in measure theory. A measurable function is defined by requiring inverse images of measurable sets to be measurable. Direct images of measurable sets need not be measurable in the codomain’s specified sigma-algebra. Accordingly, the distribution induced by a random variable \\(Z\\) is defined using inverse images:\n\n\\[\\mathbb P_Z(B)=\\mathbb P\\bigl(Z^{-1}[B]\\bigr).\\]\n\nThe image of \\(Z\\) describes its actually attained values, while its distribution assigns probabilities to measurable sets of values."
    }
  ],
  "links": [
    "mathematical-function",
    "codomain",
    "domain-of-a-function",
    "subset",
    "surjective-function",
    "real-number",
    "circle",
    "graph-of-a-function",
    "inverse-function",
    "injective-function",
    "power-set",
    "empty-set",
    "function-composition",
    "equivalence-relation",
    "bijective-function",
    "linear-map",
    "image-linear-map",
    "linear-subspace",
    "matrix",
    "matrix-rank",
    "rank-nullity-theorem",
    "homomorphism",
    "continuous-function",
    "compact-space",
    "connected-space",
    "measure-theory",
    "measurable-function",
    "probability-distribution"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}